现代电子技术
現代電子技術
현대전자기술
MODERN ELECTRONICS TECHNIQUE
2015年
11期
70-72
,共3页
高光谱遥感图像%高斯白噪声%PCA%BM3D
高光譜遙感圖像%高斯白譟聲%PCA%BM3D
고광보요감도상%고사백조성%PCA%BM3D
hyperspectral remote sensing image%Gaussian white noise%PCA%BM3D
高光谱图像光谱分辨率的提高带来数据量的显著增加，普通针对二维图像的去噪方法不能有效地应用到高光谱图像上。根据高光谱图像本身和噪声的特征，研究基于PCA的非局部去噪方法，充分利用光谱谱间相似性和谱内相似性，首先进行PCA降维选择具有代表性的维度，然后在这些维度运用非局部的BM3D方法去除噪声，最后再返回到原图像得到去噪结果。实验结果表明，该方法的去噪效果令人满意。
高光譜圖像光譜分辨率的提高帶來數據量的顯著增加，普通針對二維圖像的去譟方法不能有效地應用到高光譜圖像上。根據高光譜圖像本身和譟聲的特徵，研究基于PCA的非跼部去譟方法，充分利用光譜譜間相似性和譜內相似性，首先進行PCA降維選擇具有代錶性的維度，然後在這些維度運用非跼部的BM3D方法去除譟聲，最後再返迴到原圖像得到去譟結果。實驗結果錶明，該方法的去譟效果令人滿意。
고광보도상광보분변솔적제고대래수거량적현저증가，보통침대이유도상적거조방법불능유효지응용도고광보도상상。근거고광보도상본신화조성적특정，연구기우PCA적비국부거조방법，충분이용광보보간상사성화보내상사성，수선진행PCA강유선택구유대표성적유도，연후재저사유도운용비국부적BM3D방법거제조성，최후재반회도원도상득도거조결과。실험결과표명，해방법적거조효과령인만의。
The improvement of spectral resolution for hyperspectral images brought significant increase of data size,but the common denoising methods for two?dimensional image cannot be effectively applied to hyperspectral image. According to the characteristics of the hyperspectral image and its noise,a non?local denoising method based on principal component analysis (PCA) is studied to make full use of the similarity between the spectra and within the spectrum. Firstly,the representative di?mension is selected by reducing the dimension of PCA,and then noise is removed by the non?local BM3D method in the repre?sentative dimension. Finally,the denoising result is obtained by returning to the original image. The experiment results show that the effect of the proposed method is satisfactory.
